Business Introduction
Our client is a US-based delivery platform that operates as a service aggregator, connecting B2B business owners with a network of independent delivery providers. Their operations are spread across India, and the platform has rapidly gained traction with over 50 tenants, 1,300+ stores, and 400+ delivery partners actively using the system. In 2022 alone, the platform facilitated over 2 million unique order deliveries.
As their customer base grew, so did their need to provide transparent, actionable performance data—both to internal teams and external tenants—through structured Monthly and Quarterly Business Reviews (MBR/QBR). These reviews have traditionally been a manual process. But with scale, the client needed a solution that was automated, accurate, tenant-specific, and delivered in real-time.
Technical Background
The platform is a classic SaaS solution built using Angular and Node.js, with each tenant’s data stored in a separate MongoDB instance. The backend architecture followed strict cloud security protocols, but the data was siloed across multiple databases—one for each tenant.
As part of their review process, the client needed a consolidated reporting system that could pull in data from all tenants, transform it according to their unique KPIs, and present it through Power BI dashboards. But more than just reporting, the dashboards had to be embedded into their customer portal, offering each tenant a personalized view of their own data—while internal teams had access to the full landscape.
Business Objectives
Logesys was brought in to help the client achieve three core goals:
- Automate MBR/QBR report generation for internal and tenant use
- Create a centralized reporting database that ingests data from all tenant databases
- Embed Power BI dashboards into the customer portal with secure, role-based access
The system had to run daily, offer fresh insights without human intervention, and handle large volumes of multi-tenant data without compromising performance or cost efficiency.
Scope of Work
Our team designed and implemented an automated reporting pipeline to extract, transform, and visualize key business metrics from tenant-level MongoDBs into an Azure-based reporting ecosystem.
Challenges & Solutions
Challenge 1: Tenant Data Spread Across Separate MongoDBs
Each tenant had their own isolated MongoDB database instance. There was no built-in way to fetch data from all databases in one go, making consolidation complex and time-consuming.
Solution:
Logesys developed a robust Python extraction script that loops through all tenant connections, securely fetching the required datasets one by one. The extracted data was saved in a standardized format and served as input for the transformation pipeline. This ensured complete control over data access while maintaining scalability as new tenants are added to the platform.
Challenge 2: ADF Doesn’t Support Dynamic MongoDB Connections
Azure Data Factory (ADF) was chosen for transformation and loading—but it couldn’t natively connect to multiple MongoDB databases using dynamic parameters, making tenant-level data looping inside ADF impossible.
Solution:
We moved the looping logic entirely into Python. The Python script handled the MongoDB connections and data extraction locally on a scheduled basis. Once all tenant data was downloaded, ADF was triggered automatically to begin the transformation and load process. This hybrid approach leveraged the strengths of both tools and overcame ADF’s limitations.
Challenge 3: Performance Issues with a Large Data Model
As the final consolidated data model crossed 3 GB, the Power BI reports began experiencing performance lags, especially during interactions like filters and drilldowns.
Solution:
We restructured the data model by flattening nested structures and optimizing fact-dimensional relationships. Additionally, we reviewed and refined all DAX queries, removing unnecessary calculations and caching logic where possible. These efforts brought a noticeable boost in dashboard responsiveness without compromising the data depth.
Challenge 4: Power BI Refresh Failures on Embedded Capacity
The reports were published using Power BI Embedded (A2 tier) to control costs. But daily data refreshes began to fail due to memory limitations. Scaling up to A4 solved the issue but exceeded the budget.
Solution:
We implemented a smart workaround using ADF + Power BI APIs. ADF was configured to dynamically upscale the Power BI capacity to A4 just before the daily refresh. Once the refresh was successfully completed, the system automatically downscaled back to A2. This gave the client the performance they needed—without permanently incurring high infrastructure costs. It also ensured full automation, with no manual intervention or monitoring needed.
Value Addition
Logesys delivered a solution that not only met business objectives but also significantly enhanced the platform’s operational intelligence. Here’s how it made a difference:
- Automated MBR/QBR reports, available daily without manual effort
- Tenant-specific dashboards, securely embedded into the client portal
- Internal teams can view multi-tenant insights, aiding strategic planning
- Subscription-based access to analytics, opening a new revenue stream
- A completely cloud-native pipeline, reducing IT overhead
- Daily data refresh status emailed to the technical team automatically
- Pipelines built with flexibility to restart at any point in case of failures
Conclusion
With this implementation, the client moved from fragmented, manual business reviews to a fully automated, data-rich environment—accessible in real time by both internal teams and tenants. Logesys delivered not just a reporting system, but a foundational analytics platform that adds value every day. The client now runs smarter reviews, tracks operations with precision, and leverages analytics as a service—unlocking both performance and profit.